Exploring Human Crowd Patterns and Categorization in Video Footage for Enhanced Security and Surveillance using Computer Vision and Machine Learning

TL;DR

This paper investigates how computer vision and machine learning can analyze human crowd patterns in video footage to improve security and surveillance, focusing on motion categorization and anomaly detection.

Contribution

It introduces a novel motion categorization method using optical flow, CNNs, and machine learning to enhance scene understanding and anomaly detection in surveillance videos.

Findings

Effective motion categorization into arcs, lanes, and converging/diverging motions.

Promising accuracy in training anomaly-detection models.

Enhanced scene comprehension through behavioral insights.

Abstract

Computer vision and machine learning have brought revolutionary shifts in perception for researchers, scientists, and the general populace. Once thought to be unattainable, these technologies have achieved the seemingly impossible. Their exceptional applications in diverse fields like security, agriculture, and education are a testament to their impact. However, the full potential of computer vision remains untapped. This paper explores computer vision's potential in security and surveillance, presenting a novel approach to track motion in videos. By categorizing motion into Arcs, Lanes, Converging/Diverging, and Random/Block motions using Motion Information Images and Blockwise dominant motion data, the paper examines different optical flow techniques, CNN models, and machine learning models. Successfully achieving its objectives with promising accuracy, the results can train anomaly-detection models, provide behavioral insights based on motion, and enhance scene comprehension.